Hand motion discrimination by EMG signals without incorrect discriminations that elbow motions cause

The objective of this research is to develop a hand motion discrimination method that uses electromyogram (EMG) signals for controlling myoelectric prosthetic hands without the incorrect discriminations that elbow motions cause. We proposed a method that can discriminate hand motions using linear multiple regression models in the past. However, this method incorrectly discriminates hand motions when elbow motions are performed. The incorrect discriminations hamper myoelectric prosthetic hand use. To solve this problem, we propose a hand discrimination method using conic models. In the experiment, we prove that the proposed method can discriminate hand motions without the incorrect discriminations.